The principal goal of our proposal is an understanding of the role of reciprocal connectivity in the early visual pathway. Since the relay cells of the lateral geniculate nucleus (LGN) are the first gateway of visual input to the cortex, we shall focus on the thalamocortical loop. This focus is also an essential early step toward a long-range goal of our laboratory: the construction of a computational cortex and thalamus. Anatomical evidence of massive and ubiquitous top-down (feedback) connections in the brain has long been established. However, the role they play in the processing of visual information is largely unknown. It is impossible to distinguish by experiment between the influence of intrinsic biophysical properties and network effects on the dynamical response of neural populations, and it is therefore essential to develop theoretical and computational tools capable of probing causal effects when specific parts of the network are modified or inactivated.Our approach will entail numerical simulations of dynamically realistic neuron populations that are coupled in accord with the experimental literature. To contend with the computational burden of simulating each of the hundreds of millions of cells and their myriad connections involved on the early visual pathway, we use a statistical-based approach - the population dynamics method - that evolves probability density functions of neural states of populations. Our laboratory has demonstrated the computational efficiency of this approach in several publications. Whenever possible, we will validate our simulations with electrophysiological and optical imaging experiments conducted in Dr. Kaplan's laboratory. The model will be extended and modified so that its behavior will correspond to the experimental results.